a theory of agricultural marketing cooperatives with
Post on 22-Apr-2022
2 Views
Preview:
TRANSCRIPT
HAL Id: halshs-00906894https://halshs.archives-ouvertes.fr/halshs-00906894
Preprint submitted on 20 Nov 2013
HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.
A Theory of Agricultural Marketing Cooperatives withDirect selling
Maxime Agbo, Damien Rousselière, Julien Salanié
To cite this version:Maxime Agbo, Damien Rousselière, Julien Salanié. A Theory of Agricultural Marketing Cooperativeswith Direct selling. 2013. �halshs-00906894�
GROUPE D’ANALYSE ET DE THÉORIE ÉCONOMIQUE LYON ‐ ST ÉTIENNE
WP 1331
A Theory of Agricultural Marketing Cooperatives with Direct Selling
Maxime Agbo, Damien Rousselière, Julien Salanié
November 2013
Documents de travail | W
orking Papers
GATE Groupe d’Analyse et de Théorie Économique Lyon‐St Étienne 93, chemin des Mouilles 69130 Ecully – France
Tel. +33 (0)4 72 86 60 60
Fax +33 (0)4 72 86 60 90
6, rue Basse des Rives 42023 Saint‐Etienne cedex 02 – France
Tel. +33 (0)4 77 42 19 60
Fax. +33 (0)4 77 42 19 50
Messagerie électronique / Email : gate@gate.cnrs.fr
Téléchargement / Download : http://www.gate.cnrs.fr – Publications / Working Papers
A Theory of Agricultural Marketing Cooperatives with Direct
Selling∗
Maxime Agbo†1, Damien Rousselière‡1, and Julien Salanié§ 2,3
1Agrocampus-Ouest, Department of Economics, Management and Society, UMR GRANEM, France2Université de Lyon, Lyon, F-69007, France ; CNRS, GATE Lyon Saint-Étienne, Ecully, F-69130,
France3Université Jean Monnet, Saint-Etienne, F-42000, France
November 15, 2013
Abstract
We build a theoretical model to study a market structure of a marketing cooperative
with direct selling, in which many farmers are members of an agricultural marketing
cooperative. They can sell their production either to the cooperative or on a local market.
We show that the decision to sell to the cooperative induces an anti-competitive effect
on the direct selling market. Conversely, direct selling may create a "healthy emulation"
among farmers, leading to more production benefiting the cooperative.
Keywords: marketing cooperative, direct selling, local market, competitive.
JEL Classifications: D43; L11; Q13.
∗We thank Éric Avenel, Olivier Bonroy, Nérée Noumon, Stéphane Riou and Marc Santugini for their usefulcomments and Gemma Davies for her editorial assistance. Any remaining errors are our own. Financial supportfrom Région Pays de la Loire (France) is gratefully acknowledged.
†Corresponding author, Email: maxime.agbo@agrocampus-ouest.fr‡Email: damien.rousseliere@agrocampus-ouest.fr§Email: julien.salanie@univ-st-etienne.fr
1
1 Introduction
In this paper, we investigate a market structure in which firms simultaneously compete and
cooperate with each other. This is specifically the case of farmers around the world who are
increasingly encouraged to join marketing cooperatives in order to sell their products and open
up opportunities for direct, local market selling.
Cooperatives play an important role and hold a significant market share in agricultural
product distribution from farms to final consumers (Deller et al., 2009). For example, according
to a publication by the International Labour Office, more than 50% of global agricultural
output is marketed through cooperatives in Finland, Italy, and the Netherlands (Tchami,
2007). In 2002, agricultural cooperatives accounted for 27% of U.S. total farm marketing
expenditure (USDA, 2004). Marketing cooperatives comprise about 53% of all cooperatives
and product distribution represents 64% of the net business volume of cooperatives in the U.S.
(USDA, 2011). The rationale is that marketing cooperatives allow small farmers to get better,
or rather a secured, price by overcoming the "powerful" oligopsonist Investor-owned firms
(IOFs) (Sexton, 1990; Bontems & Fulton, 2009). With marketing cooperatives, farmers have a
much better potition for price negocation (Ladd, 1974; Cakir & Balagtas, 2012) and open up
access to markets that they cannot access individually (Camanzi et al., 2011). Cooperatives
also enable farmers to face uncertainty about agricultural market prices (Jang & Klein, 2011).
Meanwhile, there is a growing trend towards direct selling (Brown & Miller, 2008; Timmons
& Wang, 2010; Uematsu & Mishra, 2011; Low & Vogel, 2011; Fischer & Qaim, 2012). In direct
selling, consumers purchase cut out the middleman by purchasing directly from farmers. As
reported by Timmons & Wang (2010), direct food sales from farmers to consumers in the
U.S. increased by 59% from 1997 to 2007. In 2010, more than 21% of french producers were
involved in direct selling.1 Unlike cooperatives, and depending on the market conditions, with
direct selling the farmer has control over sales prices and therefore obtain greater market
power. This enables farmers to increase their profit margin (Uematsu & Mishra, 2011).2 In
practice, many farmers combine direct selling and marketing through cooperatives. They sell
part of their production to the cooperative and the rest is sold directly to the consumers. For
example in France, 19.5% of apple producers, who are members of a cooperative, sell via these
cooperatives and at the same time directly on local markets. For pear producers however, this
proportion falls to 15%.3
On the other hand, direct selling is a serious concern for some marketing cooperatives
managers. Indeed, in a context where membership and loyalty are threatened (USDA, 2002;
Bond, 2006; Bijman & Verhees, 2011; Mujawamariya et al., 2013), cooperatives increasingly
question the pertinence of allowing their members to sell directly to consumers. Specifically,
marketing cooperatives fear that they may not be able sustain the stock required for their
activity if members are engaged in direct selling. To avoid collapse, some cooperatives prevent
their members from selling through other channels than the cooperative, or settle (with mem-
1Source: Census of Agriculture, French Department of Agriculture, 2007.2According to Jang & Klein (2011), small farmers often sell their products directly to consumers for a
greater price than wholesale.3Source: Authors calculations based on the Orchard Survey, French Department of Agriculture, 2007.
2
bers) contracts which guarantee a minimum volume of delivery. For instance in France, all of
the milk production from dairy cooperative members must go through the cooperative, while
in the U.S. tobacco marketing cooperatives allow farmers to draw up contracts to distribute
their production through other channels.4 In other words, the acceptability of direct selling is
not self-evident.
The purpose of this paper is to study the market structure generated by the sale of goods
simultaneously through cooperatives and directly to consumers. Indeed, the framework of a
marketing cooperative coupled with direct selling describes a specific environment and market
structure. In many situations, farmers sell directly to consumers on local markets avoiding
transaction costs (Timmons & Wang, 2010). On the local market, farmers make decisions
independently which may lead to competition between each other. Another point to consider
is that the local market generally differs from the market where the cooperative operates.5
Therefore, for farmers, the market is composed of two parts. The first is the market on
which the marketing cooperative sells the volume of production proposed by farmers. On
this market, farmers sell collectively and cooperatively. The second is the direct local market
selling where farmers are in oligopoly. The first goal of this paper is to analyze behaviors
in the specific market structure where farmers are on the one hand in cooperation and in
oligopolistic competition on the other. We study how they share their production between
the two distribution channels, and the interdependence between the two.6 Our second goal
is to revisit cooperative managers’ concerns and see how direct selling could be beneficial for
marketing cooperatives.
We build a model in which a marketing cooperative is owned by many farmers who produce
a homogeneous good. They have the possibility to sell their production either to the coopera-
tive or on a local market. The cooperative sells farmers’ goods on a large competitive market
where the price can not be negociated. In addition to the marketing cooperative, farmers have
access to a local market where they can sell directly to consumers. We assume that the large
competitive market and the local market are separated, and the local market is oligopolistic.
The marketing cooperative does not control farmers’ production decisions, but decides the
price at which it buys the goods from farmers. In turn, farmers decide on the production
quantity and how to share this production between the local market and the cooperative.
Our analysis stresses three important results. First, we find that, under certain conditions,
farmers sell the same quantity on local market, even though they do not produce the same
amount of product. Farmers behave as if they collude together concerning quantities destined
for the local market, limiting competition, in order to get a higher price for their goods.
Compared to the standard oligopolistic structure, the presence of the marketing cooperative
allows farmers to reduce the amount of product sold on the local market, hence reducing
4We can also mention the example of New Generation Cooperatives (Cook & Chaddad, 2004; Holland &King, 2004).
5For example in some developing countries, the quantity of cocoa, coffee, cotton, pineapple, among othersmarketed through cooperatives are exported. The rest of the production is sold directly by farmers to localprocessors, final consumers, etc.
6In previous literature, economists mostly focused on the case where the two possibilities are marketing co-operatives and IOFs. In that literature, marketing cooperatives compete against IOFs (Sexton, 1990; Tennbakk,1995; Giannakas & Fulton, 2005). But in our case, the challenge for the cooperative is its own members.
3
competition on that particular market. Since they have at their disposal the opportunity to
sell through the cooperative (where they are not in competition), they seek to preserve a higher
price on the local market. We show that an increase in the number of farmers increases the
share of aggregate production delivered to the cooperative. Even though the cooperative does
not have market power in our model, it enables farmers to strengthen their market power on
the local market. Thus, the marketing cooperative plays an important role in the local market
performance. Farmers’ decision to sell to the cooperative depends on the local market demand,
the price on the large competitive market, farmers’ efficiency, and the number of farmers.
Second, we also find that, compared to a situation in which farmers do not have direct
selling opportunity, farmers deliver more to the cooperative if the maximal possible price on
local market is lower than the price on the large competitive market. Since direct selling is
operated on oligopolistic market, it brings more competitive structure to the model. In other
words, direct selling through oligopolistic market creates a "healthy emulation", and incites
farmers to produce more. But if the local demand is not sufficient enough to preserve higher
local price, this production surplus will go to the cooperative. In other words, direct selling is
not necessarily harmful for cooperatives.
The third main result concerns the case where the marketing cooperative and farmers are
selling on the same local market. We find that the two channels of distribution cannot coexist.
If the cooperative is able to purchase the goods from the farmers at a price above the market
price, then all the production is delivered to the cooperative. Such a situation could happen
if, for instance, the cooperative receives subsidies from institutions outside the market, like for
example the government. In contrast, if the cooperative cannot set its purchase price above
the market price, farmers will sell all their produce directly.
The rest of the paper is organized as follows. We present the model and derive optimal
behaviors respectively in Sections 2 and 3. Sections 4 and 5 analyze the interdependence
between direct selling performance and marketing cooperative. In Section 6, we discuss the
specific case where the marketing cooperative and farmers are operating on the same market.
We conclude in Section 7.
2 The Model
In this section, we present our model which is the hybrid case of the marketing cooperative with
direct selling. But before presenting the model, we introduce two benchmark models, which
are the two extreme cases of full marketing cooperative where the farmers have to deliver their
whole production to the cooperative, and of full direct selling where farmers are not members
of a cooperative and sell their products themselves on a local market. These two benchmark
models are used to study how the marketing cooperative with direct selling affects agents’
behavior. We also derive the quantity produced by the farmers for each benchmark.
4
2.1 Marketing cooperative with no direct selling: full marketing coopera-
tive
Consider an agricultural marketing cooperative owned by N ≥ 2 farmers. The farmers produce
a homogeneous product and sell it to the cooperative at a unit price pc. Farmer i decides and
produces the good in quantity of qi at a total cost of Ci(qi) = Aiq2i , where Ai is a positive
constant.7 In turn the cooperative sells the product on a large competitive market at a unit
price P . Here, we abstract from any processing and distribution costs borne by the cooperative.
We assume that the product delivered by farmers is sold by the cooperative without being
processed (example of fruits and vegetable, eggs, . . . , as in Tennbakk (1995)). We also assume
that the cooperative takes the price P as given, but decides on the price pc to buy the product
from the farmers.8 We consider the cooperative does not price discriminate and proposes the
same unit price to all members.9. Therefore, the profit of the cooperative is
πc = (P − pc)
N∑
j=1
qj (1)
As owner of the cooperative, each farmer receives a percentage bc ≤1N
of the profit made by
the cooperative (Karantininis & Zago, 2001).10 So, the total profit of the producer i is
ui = pcqi −Aiq2i + bcπc (2)
The game is as follows. Given the market price P , the cooperative chooses the price pc that
maximizes its profit πc. Each farmer i, leaning on pc, choose the quantity qi to produce and
to sell to the cooperative, by maximizing his profit ui.
Unlike an investor-owned firm (IOF), the cooperative’s profit maximizing is also beneficial
for the members, as a part of the profit is returned to them. Moreover, profit maximization
by the cooperative does not necessarily consist of paying farmers a low price for goods. In
order to encourage delivering, the cooperative has to decide on a reasonable price pc. In other
words, the problem of the cooperative is to choose a price pc that insures its supplying, while
maximizing the cooperation return to farmers. This consideration is similar to the seminal
work by Helmberger & Hoos (1962).
7In other words, the cooperative’s managers do not have control over farmers’ output, as in Karantininis &Zago (2001), Albaek & Schultz (1998).
8For models in which cooperatives are price-takers in their selling market, see Karantininis & Zago (2001),Albaek & Schultz (1997), Sexton (1990).
9In our case of homogeneous product, there is no reason to consider a price discriminating cooperative. Forpapers considering unique price for all farmers, see for example Albaek & Schultz (1997), Royer & Matthey(1999), Giannakas & Fulton (2005). However, a cooperative can price discriminate farmers if there are het-erogenous products of different qualities (Hendrikse & Bijman, 2002; Pascucci et al., 2012).
10Here, we assume that the farmers receive an equal part of the cooperative’s profit, as done in Tennbakk(1995). We could consider that the profit is shared proportionally to the quantity delivered by the farmer.Specifically, farmer i receives a part qi/
∑qj (like investment sharing in cooperatives, as in Albaek & Schultz
(1997)). Therefore, the profit of farmer i is ui = pcqi − Aiq2
i + µπc (qi/∑
qj) = pcqi − Aiq2
i + µ (P − pc) qi,where 0 < µ < 1 is the part of the profit the cooperative decides to share out. However, as it is easy to check,considering proportionally profit sharing would not change our results. We do a short discussion in footnote15. See also Fatas et al. (2010) for other view on benefits sharing in cooperatives.
5
The first-order condition of the farmer i yields
qi =bcP + (1− bc)pc
2Ai(3)
Using (3) the cooperative’s problem becomes11
maxpc
(P − pc)(bcP + (1− bc)pc)N∑
j=1
1
2Ai(4)
Solving the problem in (4) gives the optimal price and quantity
p̂c =(1− 2bc)P
2(1− bc)and q̂i =
P
4Ai(5)
One could remark that the farmers’ decision does not depend on bc, the part of the profit
received as owners. Actually, when making their production and delivering decision, the
farmers are more interested in the price they will be paid, as this price relies on the large
competitive price P . On the other hand, this result is also a consequence of the traditional
agricultural cooperatives’ characteristics, which is what makes them different from investor-
owned firms (IOFs). There is a trade-off between the price pc paid by the cooperative and
the profit share bc. Deriving the equilibrium price pc in (5) yields ∂p̂c∂bc
≤ 0. Farmers are
owners, controllers, and beneficiaries of the cooperative (Bijman et al., 2013). If for instance
bc decreases because of an increase of reserves, farmers (as beneficiaries) will benefit from that
increase of reserves, even though they bear a loss as owners. If bc decreases for other reasons, as
controllers the members can increase pc to compensate. In other words, the cooperative adjusts
the price pc to bc. So, only market conditions and their production efficiency matter to farmers.
Therefore, in order to succeed in preserving membership, agricultural cooperatives either need
to reinforce their market power or help farmers improve their production efficiency. Another
result worth mentioning is that, unlike Albaek & Schultz (1998), the individual production
does not depend on the number of farmers. This is because the cooperative does not have
market power and operates on a competitive market.
2.2 Direct selling without marketing cooperative: full direct selling
Consider N farmers (N ≥ 2) producing a homogeneous agricultural good that they sell to
final consumers on a local market.12 We assume that the local market is an oligopolistic one,
where the consumers demand for the agricultural good is characterized by the inverse-demand
11We consider that the cooperative maximizes its profit as considered in Hoffmann (2005), Lopez & Spreen(1985), Royer & Matthey (1999), Hovelaque et al. (2009). Boyle (2004) found empirical evidences that coopera-tives behave as if they are profit maximizers. However, some authors considered a members-welfare-maximizingcooperative (Giannakas & Fulton, 2005; Tennbakk, 1995; Sexton, 1986). Actually, as mentioned by Tennbakk(1995), there is no consensus regarding what cooperatives maximize. See Soboh et al. (2009) for more completereview on the objective functions of cooperatives.
12The definition of "local" is an ongoing debate in the literature (Hand & Martinez, 2010). In this paper, themarket is local when farmers and consumers have access with negligible transaction costs (mostly transportationcost).
6
function P o = α − βQo, Qo being the total production of all the farmers. Farmer i chooses
the quantity qoi to produce by maximizing his profit defined as
V oi =
α− β
N∑
j=1
qoj
qoi −Ai(q
oi )
2, (6)
where Ai(qoi )
2 is the production cost of i. The oligopoly production of farmer i is given as
qoi =α
(β + 2Ai)(1 + β
∑Nj=1
1β+2Aj
) . (7)
2.3 Marketing cooperative with direct selling: the hybrid case
Here, we still consider a marketing cooperative of N farmers. Each farmer i decides and
produces the good in quantity of Qi at a total cost of Ci(Qi) = AiQ2i . In addition to their
cooperative, the farmers can sell by themselves a fraction of their production directly on a
local market (as in Jang & Klein (2011) among others). Each farmer i has to choose the
proportion θi (of his production Qi) to sell to the cooperative.13 The rest 1 − θi is sold on
the local market. We assume that the local market is common for the N farmers, and is an
oligopolistic market with Cournot competition. The demand function on the local market is
PD = α− β
N∑
j=1
(1− θj)Qj
(8)
where PD is the price on the local market and α, β > 0 are parameters. The cooperative
buys the product from the farmers at the unit price Pc and sells it on a large competitive
market (non local market) at a unit price P . We assume that the competitive market where
the cooperative sells and the local market are separated.14 The profit of the cooperative is
then
Πc = (P − Pc)
N∑
j=1
θjQj (9)
13Albaek & Schultz (1997) argued that, in many cooperatives (particularly Danish cooperatives), the farmersdecide individually on the amount of production to deliver to the cooperative. Also, Giannakas & Fulton (2005)pointed out the prevalence of open-membership cooperatives, and, according to (Petraglia & Rogers, 1999),open-membership implies the cooperative will serve as quantity-taker with respect to members’ output.
14Implicitly, we consider that farmers will not decide to sell directly (i.e not through cooperative) on thenon local market. This paradigm is realistic because of transaction costs. For farmers, selling on the non localmarket would create higher transaction costs than selling on the local market. Since one of the attributes ofcooperatives is transaction costs reducing (Holloway et al., 2000; Roy & Thorat, 2008; Markelova et al., 2009;Fischer & Qaim, 2012), and farmers cannot influence the price P , it is reasonable to consider that they willprefer using the cooperative than selling directly on the non local market. We discuss in Section 6 the casewhere both the cooperative and the farmers operate on the local market.
7
As owner of the cooperative, each farmer earns a proportion bc of the profit made by the
cooperative.15 The total profit of the farmer i is16
Vi = PcθiQi −AiQ2i + bcΠc + PD (1− θi)Qi (10)
In the game, the cooperative chooses Pc that maximizes Πc. Basing on Pc, farmer j (j =
1, · · · , N) chooses Qj and θj that maximize Vj .
2.4 Remark
Even though we consider the cooperative maximizes its profits, our results cannot be gen-
eralized to investor-owned-firms (IOFs). In other words, we cannot insure that our model
remains valid beyond cooperation and direct selling environment. Indeed, IOFs will behave
like a monopsonist and will decide on and control over the quantity to purchase from the
farmers. Since farmers are price-takers, the IOF will pay them according to their supply, i.e,
aggregated marginal cost (Karantininis & Zago, 2001). In contrast, the cooperative pays ac-
cording to the conditions on the competitive market on which it sells the products delivered.
If these conditions get better, the cooperative will pay more. This is not necessary true for
the IOF who may still pay the same price, even if the price increases in the large competitive
market. In this paper, unlike IOFs, the cooperative is considered as an association for which
farmers are owners and controllers, so that farmers’ production decision coincides with (or is
close to) the cooperative’s decision. An illustration of how our results are affected is provided
in Proposition G.1 (Appendix G) where an introduction of capacity constraint changes some
of our results.
3 Optimal behaviors
In this section, we provide the optimal decision made by different agents acting in our frame-
work of cooperative with direct selling. Proposition 3.1 and 3.2 give price and quantities in
equilibrium.
15We could consider that the profit of the cooperative is shared accordingly to the quantity delivered bythe farmer, and farmer i receives a share θiQi/ΣθjQj of the profit. In that case, farmer i’s profit is PcθiQi −
AiQ2
i + b(P − Pc)θiQi + PD (1− θi)Qi, where 0 < b < 1 is the part of the profit the cooperative decides toshare out. As we said before, proportionally profit distribution does not change our results. Indeed, even if theprofit is proportionally shared, an increase of the quantity delivered by a given farmer is also profitable for theother farmers. So, proportionally profit sharing does not completely solve the free rider problem, and might notincite farmers to deliver more. This drawback of the proportionally profit sharing is highlighted by Ortmann& King (2007), regarding new members and existing members of the cooperative. On the other hand, theorganizational structure of many cooperatives is changing, and the benefits sharing rules are becoming morehybrid and complex (Deller et al., 2009; Bijman et al., 2013).
16Our way to formalize the farmer’s objective matches well with Bontems & Fulton (2009), among otherspapers. In Bontems & Fulton (2009), page 323, we can read: "The farmers who own the cooperative, are
interested in maximizing the returns from each of their own operations plus their share of the profits generated
by the cooperative".
8
Proposition 3.1. The optimal unit price at which the cooperative buys the product from the
farmers is
P ∗c =
Nα
2(1− bc)(N + β(N + 1)
∑Nj=1
12Aj
) +(1− 2bc)P
2(1− bc)(11)
Proof. See Appendix A for the proof of Propositions 3.1 and 3.2.
Proposition 3.2. The quantity produced by the farmer i is
Q∗i =
Nα
4Ai
(N + β(N + 1)
∑Nj=1
12Aj
) +P
4Ai(12)
The delivery price chosen by the cooperative is always higher in the framework of marketing
cooperative with direct selling (eq. 11) than in full marketing cooperative (eq. 5). To incite
farmers to deliver their products, the cooperative has to challenge the local market, and this
challenge becomes more serious as the conditions on the local market (α and β) allow farmers
to get higher local price PD. On the other hand, unlike the full marketing cooperative, the price
proposed by the cooperative does not necessarily decrease with bc, the part of the cooperative’s
profit earned by each farmer. Here, the result depends on the characteristics of the local market
and the large competitive market as well, i.e the challenge the cooperative has to face due to the
existence of the local market. If the challenge is not "serious" enough
(αP≤
N+β(N+1)∑
1
2Aj
N
),
then the cooperative has much leeway to adjust downward (decrease) its price Pc to an increase
of bc. On the contrary, if the challenge is too serious
(αP>
N+β(N+1)∑
1
2Aj
N
), to incite farmers
to sell to them, the cooperative has to choose a price Pc greater than P . In such a situation,
the profit is a loss, and an increase of bc implies an increase of the loss borne by members. As
a consequence, to compensate for the loss, the cooperative has to upwardly adjust the price
Pc. However, the capacity of the cooperative to price above P is limited. If the loss is too
high, as seen in Proposition 3.3, no farmer will sell to the cooperative.
Compared to the outcome of full marketing cooperating, the quantity of goods produced
by each farmer increases with the introduction of a local market. This increase in production
is driven by two facts. First, in order to compete with the local market and incite the farmer
to soll to them his whole production, the cooperative has to increase the price Pc. Second,
the increase of the production is also due to the oligopolistic structure of the local market. If
the farmers were organized in price-taker cooperative on the local market, the result would
not be the same. The market whose cooperative has better conditions would attract all the
production.
If the introduction of direct selling increases farms’s output, it is worth knowing the pro-
portion of this output farmers deliver to the cooperative.
Proposition 3.3. Let θi be the part of his production that farmer i allocates to the cooperative.
Then
. If αP≤ α then 1− θi = 0
9
. If α < αP< αi then 1− θi =
2Ai
β(N+1)
α
(N+2β(N+1)
∑1
2Aj
)−P
(N+β(N+1)
∑1
2Aj
)
Nα+P
(N+β(N+1)
∑1
2Aj
)
. If αP≥ αi then 1− θi = 1
where α =N+β(N+1)
∑1
2Aj
N+2β(N+1)∑
1
2Aj
and αi =(2Ai+β(N+1))
(N+β(N+1)
∑1
2Aj
)
2Ai
(N+2β(N+1)
∑1
2Aj
)−βN(N+1)
Proof. See Appendix B.
Proposition 3.3, illustrated in Figure 1, gives how members of the cooperative share their
production between the two markets (cooperative and local market). They compare the max-
imum price they can get on the local market (α) with the price at which the cooperative
sells on the large competitive market. Farmers are more likely to sell to the cooperative if
the other alternative (local market) is less profitable. This result is in line with the work of
Pascucci et al. (2012) concerning Italian farmers. Also, more efficient farmers (lower A) or
larger farms are more likely to sell to the cooperative. This theoretical result is confirmed by
various empirical works (Gale, 1997; Low & Vogel, 2011; Timmons & Wang, 2010; Uematsu
& Mishra, 2011). This finding also seems to conform empirical evidences found in Ethiopia
by Bernard et al. (2008).17 Moreover, using a theoretical model which is different from ours,
Jang & Klein (2011) came to similar result. Low & Vogel (2011) explain this propensity of
smaller farms for direct selling by the fact that small farms do not have enough volume to
meet the demand of large distributors, unlike large farms which have this capacity. However,
as we study in Section 4, more efficient farmers prefer to sell more to cooperative, in order to
limit competition on local market and get higher direct selling price, since they have market
power on the local market.
As we can see from Figure 1, depending on the profitability of each part (cooperative
and local market), a given farmer could sell his whole production to the cooperative and sell
nothing on the local market, or could sell his whole production on the local market and nothing
to the cooperative. Klein et al. (1997) observe this in Canadian data concerning the volume of
business members have with their cooperative. Specifically, in Proposition 3.3, α is the same
for all farmers. It means that the decision to sell or not on the local market is the same for
all the farmers. Indeed, deciding to sell on the local market depends, in general on the nature
of the product and not on the efficiency of the farmers. For instance, fruits and vegetable are
more prone to direct selling than non processed milk, rape, cocoa. If the product is valuable
and in demand on local market, each farmer perfers to distribute the goods via direct sales.
However, if the farmer decides to sell locally, the proportion of production to sell depends
on the volume of that production. The reason is that demand is not atomistic on the local
market.
17Bernard et al. (2008) compared the commercialization behavior of smallholders cooperatives members withnon cooperatives members. They found that, even though cooperatives secure higher prices for their members,smaller farmers (members) tend to reduce the output marketed through the cooperative, as a response to thehigher prices. However the opposite is true for larger farms. Of course, in our model, smaller farms (higherA) do not reduce delivering with prices increase, but their increase in selling is lower than that of larger farms(lower A).
10
0 10 20 30 40 50 60 70 80 90 1000
10
20
30
40
50
60
70
80
90
100
Price on the non local market (P )
Maxim
alpriceonthelocalmarket(α
)
Farmer i delivers all his production
to the cooperative
(θi = 1)
Farmer i sells all his production
on the local market(θi = 0) α
P= α
α
P= αi
0 < θi < 1
Figure 1: Markets characteristics and selling rate of the cooperative members.
4 The effect of the cooperative on local market transactions
After deriving the optimal decision by agents, we now start analyzing the interdependence
between the two parts (cooperative and local market) of the environment under study. Specif-
ically, in this section, we study how quantities sold and profits made by farmers on local
markets are affected by the fact that they are members of a cooperative through which they
can sell part of their production.
4.1 On sales volume
The following proposition gives a first result about quantities sold on the local market.
Proposition 4.1. If the cooperative has no capacity constraint, and αP
< αi for all i, then
all the farmers sell the same quantity of products on the local market, i.e, (1− θ1)Q1 =
(1− θ2)Q2 = · · · = (1− θN )QN .
Proof. See Appendix C.
Proposition 4.1 states that, if the market conditions are such that all farmers sell to the
cooperative, they will sell the same quantity on the local market, even if they do not have
the same cost function and do not produce the same quantity. In other words, if the local
demand is not sufficient enough to allow farmers to sell all their production on the local market
and get an interesting price, they collude on quantities. Farmers behave as if they use the
cooperative to limit competition on the local market. This is true for the higher volume
producers (lower A). The smaller farms who can sell all their production on the local market
11
do not participate in this "collusion". Indeed, as mentioned by Uematsu & Mishra (2011),
direct selling allows producers to receive a better price, because they have market power.
Therefore, farmers who produce in large quantities, sell more to the cooperative in order to
secure this higher price on local market.18 The implication is that (as one could deduce from
Proposition 3.3), farmers become more and more inclined to deliver to the cooperative, as the
number of farmers increases. And when the number of farmers increases, the cooperative can
afford to reduce the delivering price Pc, as one could see from Proposition 3.1 by considering
homogeneous production cost function. In other words, the cooperative plays an important
role in the direct selling decision. Proposition 4.2, illustrated in Figure 2 confirms this role
when the farmers have the same production cost function.
0 0.5 1 1.5 2 2.5 30
0.2
0.4
0.6
0.8
1
1.2
1.4
α
α
P= α α
P= αi
1− θi = 0 0 < 1− θi < 1 1− θi = 1
Full direct selling Coop with direct selling
Figure 2: Production total quantity sold on local market.We set N = 2, Ai = 2 ∀i, P = 1.5,
β = 0.5, and bc =1
N+1.
Proposition 4.2. Assume that the farmers have the same production cost, i.e. Aj = A, ∀j.
The total quantity sold on the local market is always lower in a marketing cooperative with
direct selling than in full direct selling, i.e.∑N
j=1 (1− θj)Q∗j ≤
∑Nj=1 q
oj .
Proof. See Appendix F.
According to Proposition 4.2, compared to the full direct selling, when farmers are orga-
nized in cooperative, they sell less on local market. This remains true even if farmers sell
nothing to the cooperative. Therefore, the possibility of selling through a cooperative enables
farmers to "protect" the local market price, which makes them better off. However, this situa-
tion could mean that consumers who purchase on the local market are worse off. This reduction
of output, due to the presence of the cooperative, is strongly stated in Proposition 4.3.
18However, as we show in Appendix G, if the cooperative is subject to a capacity constraint, and restrictsthe amount to sell, Proposition 4.1 is not valid. Therefore, farmers do not have sufficient leeway to regulateby themselves the amount of product to sell on the local market.
12
Proposition 4.3. Assume that the farmers have the same production cost, i.e. Aj = A, ∀j.
The total production of the farmers is higher in cooperation with direct selling than in full
direct selling (i.e.∑N
j=1Q∗j >
∑Nj=1 q
oj ) if α
P<
2A+β(N+1)2A
. Otherwise, the farmers produce
less in cooperation with direct selling than in full direct selling.
Proof. See Appendix E.
According to Proposition 4.3, farmers might produce less in a marketing cooperative with
direct selling than in full direct selling, if local demand is such that they can get a better price
on the local market. In other words, it is possible that farmers produce less when they have
access to two markets (cooperative and local market) than when they have access to just one
market (local market). From this result and the fact that farmers always produce more in a
marketing cooperative with direct selling than in a full marketing cooperative, we suspect that
the presence of the cooperative might be a source of output reduction. Indeed, it is well known
in the literature that a cooperative always produces more output than the investor-owned firm
(Nourse, 1922; Tennbakk, 1995; Bontems & Fulton, 2009; Giannakas & Fulton, 2005). Many
authors made the case for the pro-competition effect of the cooperative (Petraglia & Rogers,
1999; Deller et al., 2009)19. Specifically, there is a pro-competitive effect in the case where the
cooperative does not have control over farmers’ production decisions. If the cooperative can
restrict the farms’ product then it might reduce output (Youde, 1978; Sexton & Iskow, 1988;
Sexton, 1990). However, in our model, the cooperative does not have control over farmers’
production, and yet, it might reduce farms’ output. In fact, what we know from the literature
is that a cooperative induces a competition effect when challenging an IOF. But little research
has been carried out concerning the situation of marketing cooperatives challenging their own
members who operate on an oligopolistic local market. Indeed, marketing cooperatives can
reduce output for two reasons. First, according to Propositions 3.3 and 4.3, a cooperative
reduces farms’ output, in the case where no farmers sell to the cooperative. Without the
cooperative, farmers would be in a standard oligopolistic market. With the cooperative, since
farmers know that they have another option (selling to cooperative), they can increase price on
the local market by reducing output. As we will see, this decision to reduce output, increases
farmers profits. The second reason is that our framework is a combination of cooperation and
oligopolistic competition. Within a certain context, it is well known that the cooperative game
structure yields less output than oligopolistic structure. Therefore, the combination of both
structures may reduce output, if the performance of the cooperative is not sufficient enough
to incite farmers to sell.
4.2 On farmers’ profits
In the preceding section, we have seen that with a marketing cooperative, farmers reduce their
sales on the local market. We now study how this result affects the individual profit made by
farmers on the local market. Specifically, we compare the profit made from selling through
full direct selling with the profit made on local markets through a marketing cooperative with
19Sexton (1990) used the term "yardstick of competition".
13
direct selling. In general, the result is ambiguous. We therefore use graphical analysis to find
that the local market profit might be higher, depending on the parameters of the model. In
Figure 3, we plot the profits of the farmers in full direct selling and in a marketing cooperative
with direct selling, as functions of the number of farmers. For few farmers, the profit made on
0 50 100 150 200 250 3000
0.05
0.1
0.15
0.2
0.25
N
Farmer i profit (on local market) in full direct selling Farmer i profit on local market in marketing cooperative with direct selling
Figure 3: Farmers’ individual profits on local market. We set Ai = 2 ∀i, P = 1.15, α = 1, β = 0.1,
and bc =1
N+1.
local markets is higher in full direct selling than in marketing cooperative with direct selling.
But as the number of farmers (N) goes up, competition also increases on the local market.
Unlike the framework of a marketing cooperative with direct selling, in full direct selling farmers
do not have the opportunity to turn to other markets and to limit competition on the local
market. As a consequence, the profit decreases faster in full direct selling than in marketing
cooperative with direct selling. Above a certain value of N , the profit in marketing cooperative
with direct selling becomes higher.
It is well known that, in standard oligopolistic market, the firms’ individual profit decreases
with the number of firms, as we see in Figure 3 for the farmers’ individual profit in full direct
selling. However, the result is not necessarily the same when the oligopolistic firms are also
organized in a marketing cooperative through which they can sell their production on other
markets. Indeed in our framework, for some characteristics of the markets, individual profit
may increase with the number of farmers. As shown in Figure 3, for lower values of N , an
increase of N leads the farmers to sell more to the cooperative in order to avoid a price
decrease on the local market. Under certain conditions, the positive effect (of selling to the
cooperative) on the local market profit may overcompensate for the negative effect of an
increase of the number of farmers. As a consequence, the individual profit could increase. But
for many farmers, the competition on local market becomes too hard to be compensated for.
In Appendix H, we draw the profits as functions of α, β, P , and A.
14
Proposition 4.4. Let V Di be the profit of the farmer i made on the local market, in a structure
of marketing cooperative with direct selling. We denote by K the quantity of goods that the
farmer decides to sell on local market. Then,
If K >α
2(N − 1)βthen
∂V Di
∂K< 0
Proof. See Appendix D.
Proposition 4.4 gives a more formal result about the effect selling has on the individual
farmer’s profit made on the local market. If, depending on his efficiency, a farmer produces a
greater volume of goods than a specific threshold, then it is better for him to sell part of this
production to the cooperative. In other words, even if direct selling opportunity exists, large
farm holders may still have a vested interest in being members of a marketing cooperative.
In this section, we have seen that the volume of business farmers conduct with their
marketing cooperative, affects their performance in direct selling. Even if the cooperative
does not have market power, it enables farmers to strengthen their market power on the local
market. However, this role of the marketing cooperative is undermined if, for some reasons,
the cooperative can restrict the amount of product farmers are allowed to sell. Such a situation
could occur if the cooperative has a capacity constraint or the decisions of the farmers does
not correspond to those made by the managers of the cooperative.
5 Direct selling and delivery to the cooperative
In this section, we examine the effect of the introduction of direct selling on the quantities
delivered by farmers to their cooperative. Indeed, there are some concerns about whether the
cooperative should allow or not its members to sell their production by themselves directly to
the consumer. The legitimacy of these concerns lies in the fact that the marketing cooperative
that allows its members to sell directly on a local market is likely to experience supply shortage.
However, we show that allowing direct selling might be beneficial for the cooperative.
Proposition 5.1. Let θj be the fraction of his production Q∗j that a farmer j (j = 1, · · · , N)
allocates to the cooperative, in a model of cooperative with direct selling. We denote by q̂j the
production of farmer j in a model of cooperative without direct selling. Then,
If α < P then∑N
j=1 θjQ∗j >
∑Nj=1 q̂j
Proof. The proof consists simply in calculating the expression of each side of the inequality
and making the comparison.
Proposition 5.1 (illustrated by Figure 4) states that, if the maximum price farmers get on
the local market is lower than the price the cooperative gets on the large competitive market,
then the farmers will sell more to the cooperative in marketing cooperative with direct selling
than in full marketing cooperative. Introducing direct selling brings two modifications to the
initial full marketing cooperative. First, in addition to the cooperative, farmers have access
to other market. In other words, they have more opportunity to sell their product. Second,
15
0 0.5 1 1.5 2 2.5 30
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
α
α
P= α α
P= αiα = P
Full marketing cooperative Coop with direct selling
Figure 4: Production total quantity delivered to the cooperative.We set N = 2, Ai = 2 ∀i,
P = 1.5, β = 0.5, and bc =1
N+1.
direct selling adds an oligopoly structure to the initial cooperation. With oligopoly structure,
farmers compete with each other more than in cooperation. These two modifications incite
farmers to produce more. But if they produce more and the local demand is not sufficient to
generate a reasonnable price, the cooperative will be sold more produce. Direct selling may
create a healthy emulation that could be beneficial for the cooperative. In contrast, if demand
on the local market is promising then the cooperative will be worse off. Therefore, the decision
of whether a cooperative should allow direct selling depends on the characteristics of the local
market, regarding the competitive price of the large market.
As shown in Figure 5 (5a, 5b, and 5d), if α is lower than P , the consequence of direct
selling might be an increase of the profit made by the farmers from their business with the
cooperative. It is the straightforward implication of the fact that farmers provide more and
the cooperative proposes a higher price. In other words, the healthy emulation leading to
more production, might have a positive effect on the performance of farmers’ activities with
the cooperative.
In total, if for some given values of parameters, direct selling has a positive effect on
individual profit made with cooperative, and the cooperative has a positive effect on individual
profit made on the local market, then a marketing cooperative with direct selling can have
positive effect on total profit (sum of profit made on the local market and profit made with the
cooperative). We give an illustration in Figure 6 where we draw the total individual profits of
farmers in the three market structures considered in this paper.
In the preceding sections (4 and 5), we have studied the interdependence between the
cooperative market and the direct selling market. Specifically, we have seen how quantities
produced and sold are affected by each part. We now summarize these results in Figure 7
16
0 0.5 1 1.50
0.05
0.1
0.15
0.2
0.25
α
Figure 5a
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
0.1
0.2
0.3
0.4
0.5
0.6
0.7
P
Figure 5b
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
x 10−3
0
50
100
150
A
Figure 5c
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.2
0.25
0.3
0.35
β
Figure 5d
Full marketing cooperative Coop with direct selling
Figure 5: Farmers’ individual profits from their activities with the cooperative. For Figure 5a,
we set N = 2, Ai = 1 ∀i, P = 1, β = 0.2, and bc = 1
N+1. For Figure 5b, N = 2, Ai = 1 ∀i, α = 1,
β = 0.7, and bc =1
N+1. For Figure 5c, N = 2, P = 0.5, α = 1.5, β = 0.01, and bc =
1
N+1. For Figure
5d, N = 2, Ai = 1 ∀i, α = 1, P = 1.15, and bc =1
N+1
which gives quantities produced and sold on the two parts, under different market structures.
6 Discussion: The case where cooperative and members operate
on the same local market
Up to now, we have considered that the direct selling market and the market where the
cooperative sells are separate, in that farmers do not sell directly on the large competitive
market, and the cooperative does not operate on the local market. Due to transaction costs,
it is reasonable to assume that farmers will not decide to sell on the large competitive market.
In contrast, the marketing cooperative can decide to sell on the local market at the same time.
In this section, we study what happens if farmers and the cooperative decide to be present
on the local market. We show that, in such a situation, direct and cooperative selling can no
longer coexist.
Indeed, the farmer i’s total profit is V Li = PcθiQi − AiQ
2i + bcΠc + PD (1− θi)Qi where
Πc = (PD − Pc)∑N
j=1 θjQj and PD = α − β(∑N
j=1Qj
). The first derivative of V L
i with
respect to θi is equal to (1−bc)(Pc − α+ β
(∑Nj=1Qj
))Qi. There are three cases: Pc < PD,
Pc > PD, and Pc = PD. If Pc < PD then farmers will not sell to the cooperative, i.e. θj = 0
for all j, and there is no cooperative selling and only direct selling. The cooperative would
disappear. If Pc > PD then farmers will sell all their production to the cooperative, i.e. θj = 1
for all j and there is no direct selling. Such a situation could occur if the cooperative receives
17
0 0.5 1 1.50
0.1
0.2
0.3
0.4
0.5
α
Figure 6a
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
0.1
0.2
0.3
0.4
0.5
0.6
0.7
P
Figure 6b
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
x 10−3
0
50
100
150
200
A
Figure 6c
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
β
Figure 6d
Full direct selling Full marketing cooperative Coop with direct selling
Figure 6: Farmers’ individual total profits for different market structures. For Figure 6a, we set
N = 2, Ai = 1 ∀i, P = 1, β = 0.2, and bc = 1
N+1. For Figure 6b, N = 2, Ai = 1 ∀i, α = 1, β = 0.7,
and bc = 1
N+1. For Figure 6c, N = 2, P = 0.5, α = 1.5, β = 0.01, and bc = 1
N+1. For Figure 6d,
N = 2, Ai = 1 ∀i, α = 1, P = 1.15, and bc =1
N+1
α
qoi
P4Ai
q̂i
Q∗i
α = αP α = αiP
1− θi = 0 0 < 1− θi < 1 θi = 0
α = P
farmer i’s delivery in a structure of coop with direct sellingfarmer’s i’s direct sales in a structure of coop with direct selling
Figure 7: Quantities produced and delivery schedule
18
subsidies from the government or other institutions outside the market. Suppose now that
Pc = PD, i.e. Pc − α+ β(∑N
j=1Qj
)= 0. The first-order condition with respect to Qi yields
the optimal quantity QLi produced by the farmer i as
QLi =
α
(2Ai + β(1− θi))(1 + β
∑Nj=1
1+bcθj2Aj+β(1−θj)
) (13)
Since Pc = PD, the profit of the cooperative is equal to zero and the total profit of any
member i becomes V Li =
(α− β
(∑Nj=1Qj
))Qi−AiQ
2i , which is maximized by the oligopoly
production.20 As a consequence, we have θi = 0, i.e. there is nothing sold to the cooperative
and only direct selling.
In other words, direct selling is not suitable for the cooperatives who sell on the same
market as farmers. It is more appropriate for large cooperatives who can have access to large
foreign markets.
7 Concluding Remarks
In this paper, we study the market structure of a marketing cooperative with direct selling, in
which many farmers organized in cooperatives, can sell both to the cooperative and directly
to consumers on a local market. Using a theoretical model, we show that direct selling and
marketing through a cooperation might be beneficial for each party. Even though the cooper-
ative does not have market power, it could have an anti-competitive effect on the direct selling
market, and thus benefit farmers. Conversely, direct selling can create a healthy emulation
and incite farmers to increase production that could benefit the cooperative.
The contribution of this paper is twofold. First, it is an attempt to build a theoretical
framework of a marketing cooperative coupled with direct selling. To our knowledge, this is
among the first theoretical works about direct selling and cooperatives. Jang & Klein (2011)
also considered a framework in which farmers simultaneously sell to marketing cooperatives
and on local markets. But in Jang & Klein (2011), prices and demand are given. Farmers
only decide on how to share production between the two parts. Second, this paper contributes
to clear up the cooperatives’ concerns about whether to allow members to directly sell to
consumers.
There are several ways to extend our paper. We have considered that farmers produce a
homogeneous unique good. We have also abstracted from the possibility for farmers to sell to
IOFs. However, in many cases, they produce more than one good, and decide on which one to
sell to the cooperative, through direct selling, or to IOFs. Moreover, cooperatives can demand
a minimum volume (fixed volume or percentage of production) from its members. A minimum
supply might mitigate the anti-competitive effect we mention in this paper. Finally, it might
be interesting to go over the standard oligopolistic setup we consider for the local market, as
it is well known that direct selling could involve spatial competition.
20The oligopoly production is the quantity produced by the farmers when there is no cooperative and thefarmers are in oligopolistic competition on the local market. We found in Section 2.2 the expression of theoligopoly production.
19
References
Albaek, S., & Schultz, C. 1997. One Cow, One Vote? Scandinavian Journal of Economics,
99(4), 597–615.
Albaek, S., & Schultz, C. 1998. On the Relative Advantage of Cooperatives. Economics
Letters, 59(3), 397–401.
Bernard, T., Taffesse, A. S., & Gabre-Madhin, E. 2008. Impact of Cooperatives on Smallhold-
ers’ Commercialization Behavior Evidence from Ethiopia. Agricultural Economics, 39(2),
147–161.
Bijman, J., & Verhees, F. 2011. Members or Customers? Farmers Commitment to Sup-
ply Cooperatives. Paper presented at the International Conference on the Economics and
Management of Networks, 1–3 December 2011, Limassol, Cyprus.
Bijman, J., Hendrikse, G., & van Oijen, A. 2013. Accommodating Two Worlds in One Or-
ganisation: Changing Board Models in Agricultural Cooperatives. Managerial and Decision
Economics, 34(3-5), 204–217.
Bond, J. K. 2006. Cooperative Growth and Decline: A Game Theoretic Approach to Un-
derstanding Members Allocation Choices. Selected Paper prepared for presentation at the
American Agricultural Economics Association Annual Meeting, Long Beach, California, July
23-26, 2006, available at http://ageconsearch.umn.edu/bitstream/21279/1/sp06ke04.pdf .
Bontems, P., & Fulton, M. 2009. Organizational Structure, Redistribution and the Endogeneity
of Cost: Cooperatives, Investor-owned Firms and the Cost of Procurement. Journal of
Economic Behavior and Organization, 72(1), 322–343.
Boyle, G. E. 2004. The Economic Efficiency of Irish Dairy Marketing Cooperatives. Agribusi-
ness, 20(2), 143–153.
Brown, C., & Miller, S. 2008. The Impacts of Local Markets: A Review of Research on
Farmers Markets and Community Supported Agriculture. American Journal of Agricultural
Economics, 90(5), 1296–1302.
Cakir, M., & Balagtas, J. V. 2012. Estimating Market Power of U.S. Dairy Cooperatives in
the Fluid Milk. American Journal of Agricultural Economics, 94(3), 647–658.
Camanzi, L., Malorgio, G., & Garcia Azcarate, T. 2011. The role of Producer Organizations
in Supply Concentration and Marketing: A comparison between European Countries in the
Fruit and Vegetable Sector. Journal of Food Products Marketing, 17(2–3), 327–354.
Cook, M. L., & Chaddad, F. 2004. Redesigning Cooperative Boundaries: the Emergence of
New Models. American Journal of Agricultural Economics, 86(5), 1249–1253.
Deacon, Robert T. 2012. Fishery management by harvester cooperatives. Review of Environ-
mental Economics and Policy, 6(2), 258–277.
20
Deller, S., Hoyt, A., Hueth, B., & Sundaram-Stukel, R. 2009. Research on the Economic
Impact of Cooperatives. University of Wisconsin Center for Cooperatives.
Fatas, E., Jimenez-Jimenez, F., & Morales, A. J. 2010. Blind Fines in Cooperatives. Applied
Economic Perspectives and Policy, 32(4), 564–587.
Fischer, E., & Qaim, M. 2012. Linking Smallholders to Markets: Determinants and Impacts
of Farmer Collection Action in Kenya. World Development, 40(6), 1255–1268.
Gale, F. 1997. Direct Farm Marketing as a Rural Development Tool. Rural Development
Perspectives, United States Department of Agriculture, 12(2), 19–25.
Giannakas, K., & Fulton, M. 2005. Process Innovation Activity in a Mixed Oligopoly: The
Role of Cooperatives. American Journal of Agricultural Economics, 87(2), 406–422.
Hand, M. S., & Martinez, S. 2010. Just What Does Local Mean. Choices, 25(1), Available at
http://farmdoc.illinois.edu/policy/choices/20101/2010102/2010102.pdf.
Helmberger, P., & Hoos, S. 1962. Cooperative Enterprise and Organization Theory. Journal
of Farm Economics, 44(2), 275–290.
Hendrikse, G. W. J., & Bijman, J. 2002. On the Emergence of New Growers’ Associations:
Self-selection versus Market Power. European Review of Agricultural Economics, 29(2),
255–269.
Hoffmann, R. 2005. Ownership Structure and Endogenous Quality Choice: Cooperative vs
Investor Owned Firms. Journal of Agricultural and Food Industrial Organization, 3(2), 1–24.
Holland, S. J., & King, R. P. 2004. Trading Mechanisms for New Generation Cooperative
Stock: the Architecture of Organizational Formation and Demise. American Journal of
Agricultural Economics, 86(5), 1262–1268.
Holloway, G., Nicholson, C., Delgado, C., Staal, S., & Ehui, S. 2000. Agroindustrialization
through Institutional Innovation: Transaction Costs, Cooperatives and Milk-market Devel-
opment in the East-African Highlands. Agricultural Economics, 23(3), 279–288.
Hovelaque, V., Duvaleix-Treguer, S., & Cordier, J. 2009. Effects of Constrained Supply and
Price Contracts on Agricultural Cooperatives. European Journal of Operational Research,
199(3), 769–780.
Jang, W., & Klein, C. M. 2011. Supply Chain Models for Small Agricultural Enterprises.
Annals of Operations Research, 190(1), 359–374.
Karantininis, K., & Zago, A. 2001. Endogenous Membership in Mixed Duopsonies. American
Journal of Agricultural Economics, 83(5), 1266–1272.
Klein, K. K., Richards, T. J., & Walburger, A. 1997. Determinants of Cooperative patronage
in Alberta. Canadian Journal of Agricultural Economics, 45(2), 93–110.
21
Ladd, G. W. 1974. A Model of Bargaining Cooperative. American Journal of Agricultural
Economics, 56(3), 509–519.
Lopez, R. A., & Spreen, T. H. 1985. Co-ordination Strategies and Non-members’ Trade in
Processing Cooperatives. Journal of Agricultural Economics, 36(3), 385–396.
Low, S. A., & Vogel, S. 2011. Direct and Intermediated Marketing of Local Foods in the United
States. ERR - 128, U.S, Department of Agriculture.
Markelova, H., Meinzen-Dick, R., Hellin, J., & Dohrn, S. 2009. Collective Action for Small-
holder Market Access. Food Policy, 34(1), 1–7.
Mujawamariya, G., D’Haese, M., & Speelman, S. 2013. Exploring Double Side-Selling in
Cooperatives, Case Study of Four Coffee Cooperatives in Rwanda. Food Policy, 39(1),
72–83.
Nourse, E. G. 1922. The Economic Philosophy of Co-operation. American Economic Review,
12(4), 577–597.
Ortmann, G.F., & King, R. P. 2007. Agricultural Cooperatives I: History, Theory and Prob-
lems. Agrekon, 46(1), 40–68.
Pascucci, S., Gardebroek, C., & Dries, L. 2012. Some Like to Join, Others to Deliver: An
Econometric Analysis of Farmers’ Relationship with Agricultural Cooperatives. European
Review of Agricultural Economics, 39(1), 51–74.
Petraglia, L. M., & Rogers, R. T. 1999. The Impact of Agriculture Marketing Cooperatives on
Market Performance in U.S. Food Manufacturing Industries for 1982. Research Report No.
12, Food Marketing Policy Center, The University of Connecticut.
Roy, D., & Thorat, A. 2008. Success in High Value Horticultural Export Markets for the Small
Farmers: The Case of Mahagrapes in India. World Development, 36(10), 1874–1890.
Royer, J. S., & Matthey, H. 1999. Optimal Strategies of Marketing Cooperatives Regarding Non-
members Business. Working Paper. Selected For the 1999 Annual meeting by the American
Agricultural Economics Association, August 8-11, Nashville, TN .
Sexton, R. J. 1986. The Formation of Cooperatives: A Game-Theoretical Approach with
Implications for Cooperative Finance, Decision Making, and Stability. American Journal of
Agricultural Economics, 68(2), 214–225.
Sexton, R. J. 1990. Imperfect Competition in Agricultural Markets and the Role of Coopera-
tives: A Spatial Analysis. American Journal of Agricultural Economics, 72(3), 709–720.
Sexton, R. J., & Iskow, J. 1988. Factors Critical to the Success or Failure of Emerging Agri-
cultural Cooperatives. Giannini Foundation Information Series No. 88-3.
22
Soboh, R. A. M. E., Lansink, A. O., Giesen, G., & van Dijk, G. 2009. Performance Measure-
ment of the Agricultural Marketing Cooperatives: The Gap between Theory and Practice.
Review of Agricultural Economics, 31(3), 446–469.
Tchami, G. 2007. Handbook on Cooperatives for Use by Worker’s Organizations. Geneva:
International Labour Office.
Tennbakk, B. 1995. Marketing Cooperatives in Mixed Duopolies. Journal of Agricultural
Economics, 46(1), 33–45.
Timmons, D., & Wang, Q. 2010. Direct Food Sales in the United States: Evidence from State
and County-Level Data. Journal of Sustainable Agriculture, 34(2), 229–240.
Uematsu, H., & Mishra, A. K. 2011. Use of Direct Marketing Strategies by Farmers and their
Impact on Farm Business Income. Agricultural and Resource Economics Review, 40(1),
1–19.
USDA. 2002. Problems and Issues Facing Farmers Cooperatives. Rural Business-Cooperative
Service, RBS Research Report 192, United States Department of Agriculture.
USDA. 2004. Farm Marketing, Supply and Service Cooperative Historical Statistics. Cooper-
ative Information Report 1, Section 26, United States Department of Agriculture.
USDA. 2011. Cooperative Statistics, 2010. Service Report 71, United States Department of
Agriculture.
Youde, J. G. 1978. Cooperative Membership Policies and Market Power. Agricultural Coop-
eratives and the Public Interest. North Central Regional Research Project N.C. 117 Studies
of the Organization and Control of the U.S. Food System Monograph., 4, 219–225.
23
A Proof of Propositions 3.1 and 3.2
The farmer i’s problem is
maxQi,θi
Vi = PcθiQi −AiQ
2i + bc(P − Pc)
N∑
j=1
θjQj
+ α (1− θi)Qi − β (1− θi)Qi
N∑
j=1
(1− θj)Qj
. (14)
Writing the first-order condition with respect to θi and simplifying by Qi yields
Pc + bc(P − Pc)− α+ β
N∑
j=1
(1− θj)Qj + (1− θi)Qi
= 0. (15)
From (15), (1− θi)Qi =α− bcP − (1− bc)Pc
β−
N∑
j=1
(1− θj)Qj . (16)
Considering equation (16) for each producer, we have
(1− θ1)Q1 = (1− θ2)Q2 = · · · = (1− θN )QN . (17)
Therefore equation (16) gives Qi =α− bcP − (1− bc)Pc
β(N + 1) (1− θi). (18)
The first-order condition with respect to Qi is
Pcθi − 2AiQi + bc(P − Pc)θi + α (1− θi)
−β
(1− θi)
N∑
j=1
(1− θj)Qj + (1− θi)2Qi
= 0. (19)
Or Pcθi − Pc + Pc − 2AiQi + bc(P − Pc)θi − bc(P − Pc) + bc(P − Pc)
+α (1− θi)− β
(1− θi)
N∑
j=1
(1− θj)Qj + (1− θi)2Qi
= 0. (20)
So, Pc − (1− θi)Pc − 2AiQi + bc(P − Pc)− bc(P − Pc) (1− θi)
+α (1− θi)− β (1− θi)
N∑
j=1
(1− θj)Qj + (1− θi)Qi
= 0. (21)
24
Let K ≡α− bcP − (1− bc)Pc
β(N + 1); So, ∀j = 1, · · · , N, (1− θj)Qj = K. (22)
Rearranging (21) and using (22), we find
1− θi =2AiK
bcP + (1− bc)Pc. (23)
Let us turn now to the cooperative’s problem which is
maxPc
Πc = (P − Pc)N∑
j=1
θjQj . (24)
Using (18), (22), and (23), one can find that
N∑
j=1
θjQj = (bcP + (1− bc)Pc)N∑
j=1
1
2Ai−NK. (25)
The problem of the cooperative becomes
maxPc
Πc = (P − Pc)
(bcP + (1− bc)Pc)
N∑
j=1
1
2Ai−NK
. (26)
Deriving the first-order condition yields the expected results for P ∗c and Q∗
i .21
B Proof of Proposition 3.3
From (22) and (23) we have 1− θi =2Ai
β(N+1)
[α
bcP+(1−bc)Pc− 1
]We find the result by plugging
the expression of Pc into the expression above for 1− θi, and taking account of the fact that
θi has to be between 0 and 1.
C Proof of Proposition 4.1
The problem of the farmer i is
maxQi,θi
Vi = PcθiQi −AiQ
2i + bc(P − Pc)
N∑
j=1
θjQj
+ α (1− θi)Qi − β (1− θi)Qi
N∑
j=1
(1− θj)Qj
, (27)
subject to the constraint 0 ≤ θi ≤ 1. If αP
< αi for all i, then the constraint θi ≥ 0 is not
bounded. From Equation (22) we find the result.
21The second-order condition confirms these results.
25
D Proof of Proposition 4.4
The profit made by the farmer i on the local market is V Di = PD (1− θi)Q
∗i−(1− θi)Ai(Q
∗i )
2.22
We can write V Di = (α−NβK)K−AiKQ∗
i . Using (18), (22) and (23), we can find that Q∗i =
bcP+(1−bc)Pc
2Ai= α−β(N+1)K
2Ai. We can rewrite and rearrange V D
i as V Di = 1
2
(αK − (N − 1)βK2
).
We can then calculate the first derivative of V Di with respect to K and get to the result.
E Proof of Proposition 4.3
If Aj = A, ∀j, then∑N
j=1Q∗j = Nα
4A+2β(N+1) +NP4A and
∑Nj=1 q
oj = Nα
2A+β(N+1) . We can then
compare∑N
j=1Q∗j and
∑Nj=1 q
oj .
F Proof of Proposition 4.2
We assume that Aj = A, ∀j. Basing on Proposition 3.3, we distinguish three cases. IfαP≤ α = 2A+β(N+1)
2A+2β(N+1) then 1− θj = 0 for all j and we get to the result. If αP≥ α = 2A+β(N+1)
2A
then 1 − θj = 1 for all j, and∑N
j=1 (1− θj)Q∗j =
∑Nj=1Q
∗j ≤
∑Nj=1 q
oj , from Proposition 4.3.
If α < αP< α then
∑Nj=1 (1− θj)Q
∗j −
∑Nj=1 q
oj = Nα(2A+2β(N+1))
β(N+1)(4A+2β(N+1) −NP
2β(N+1) −Nα
2A+β(N+1) =
N2β(N+1)
[2Aα
2A+β(N+1) − P]< 0.
G Cooperative with capacity constraint
In this paper, we assume that the cooperative accepts to purchase the entire quantity of
products the farmers decide to sell. However, sometimes, the cooperative is subject to a
capacity constraint (defective storage facilities, governmental quotas, etc.) and cannot buy the
whole volume for sale. The capacity constraint K̄ is exogeneous and known by the cooperative
members. For example, it is the case of harvesters cooperatives who manage collective fishing
quotas (Deacon, 2012). Proposition G.1 reexamines Proposition 4.1 within the context of
restriction implied by capacity constraint.
Proposition G.1. If the cooperative has a capacity constraint, then the farmers do not nec-
essary sell the same quantity on the local market, as seen in Proposition 4.1.
Proof. Let us denote by K̄ the maximal amount of products the cooperative can buy from the
22We assume that the production cost is shared over the local market and the large market proportionally tothe percentage of production sold on each market. This way to share cost is inspired by Karantininis & Zago(2001).
26
farmers. The farmer i’s problem is
maxQi,θi
Vi = PcθiQi −AiQ
2i + bc(P − Pc)
N∑
j=1
θjQj
+ α (1− θi)Qi − β (1− θi)Qi
N∑
j=1
(1− θj)Qj
subject to
N∑
j=1
θjQj ≤ K̄. (28)
Writing the first-order condition with respect to θi and simplifying by Qi yields
Pc + bc(P − Pc)− α+ β
N∑
j=1
(1− θj)Qj + (1− θi)Qi
− λi = 0, (29)
where λi ≥ 0 is the Lagrange multiplier corresponding to the capacity constraint for the
problem in equation (28).
From (29), β (1− θi)Qi − λi = α− bcP − (1− bc)Pc − β
N∑
j=1
(1− θj)Qj . (30)
Considering equation (30) for each producer, we have
(1− θ1)Q1 −λ1
β= (1− θ2)Q2 −
λ2
β= · · · = (1− θN )QN −
λN
β. (31)
From (31), if the capacity constraint is not bounded, then the farmers sell an equal quantity
on the local market. Assume now that the capacity constraint is bounded(∑N
j=1 θjQj = K̄).
The first-order condition with respect to Qi is
Pcθi − 2AiQi + bc(P − Pc)θi + α (1− θi)
−β
(1− θi)
N∑
j=1
(1− θj)Qj + (1− θi)2Qi
− λiθi = 0. (32)
From (29), λi = Pc + bc(P − Pc) − α + β[∑N
j=1(1− θj)Qj + (1− θi)Qi
]. Therefore, (32)
becomes
2AiQi + β (1− θi)Qi = α− β
N∑
j=1
(1− θj)Qj . (33)
Considering (33) for each farmer, we get to
(1− θ1)Q1 +2A1Q1
β= (1− θ2)Q2 +
2A2Q2
β= · · · = (1− θN )QN +
2ANQN
β. (34)
27
We now need to prove that the AjQj may vary with j when there is capacity constraint.
An example is the case of direct selling with no marketing cooperative (K̄ = 0) studied in
Section 2.2.
If there is an increase of K̄ then the λjs will decrease (λi is the shadow price of the
cooperative’s capacity for farmer i). Therefore, due (29), the quantity∑N
j=1(1 − θj)Qj +
(1− θi)Qi will decrease, which means the farmers will decrease their sales amount on the
local market.
H Figures
0 0.5 1 1.50
0.1
0.2
0.3
0.4
0.5
α
Figure 8a
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
0.02
0.04
0.06
0.08
0.1
0.12
P
Figure 8b
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.045
10
15
20
25
A
Figure 8c
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.05
0.1
0.15
0.2
0.25
β
Figure 8d
Full direct selling Coop with direct selling
Figure 8: Farmers’ individual profits made on local market. For Figure 8a, we set N = 2, Ai = 1
∀i, P = 1, β = 0.2, and bc = 1
N+1. For Figure 8b, N = 2, Ai = 1 ∀i, α = 1, β = 0.7, and bc = 1
N+1.
For Figure 8c, N = 2, P = 0.5, α = 1.5, β = 0.01, and bc = 1
N+1. For Figure 8d, N = 2, Ai = 1 ∀i,
α = 1, P = 1.15, and bc =1
N+1
28
0 10 20 30 40 50 60 70 80 90 1000
2
4
6
8
10
12
14
16
18
20
N
Full marketing cooperative Coop with direct selling
Figure 9: The profits made by the cooperative. We set P = 1, Ai = 1 ∀i, α = 1, β = 0.5, and
bc =1
N+1.
0 0.5 1 1.50
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
α
Figure 10a
19 19.1 19.2 19.3 19.4 19.5 19.6 19.7 19.8 19.9 20175
180
185
190
195
200
205
P
Figure 10b
10 11 12 13 14 15 16 17 18 19 200.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
A
Figure 10c
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 51.5
1.55
1.6
1.65
1.7
1.75
1.8
1.85
β
Figure 10d
Full marketing cooperative Coop with direct selling
Figure 10: The profits made by the cooperative. For Figure 10a, we set N = 2, Ai = 1 ∀i, P = 1,
β = 0.5, and bc =1
N+1. For Figure 10b, N = 2, Ai = 1 ∀i, α = 5, β = 0.5, and bc =
1
N+1. For Figure
10c, N = 2, P = 1.5, α = 1, β = 0.5, and bc =1
N+1. For Figure 10d, N = 2, Ai = 1 ∀i, α = 1, P = 2,
and bc =1
N+1
29
top related